Abstract
In prostate cancer, ERF loss-of-function mutations enhance ERG activity to promote tumorigenesis.
Major finding: In prostate cancer, ERF loss-of-function mutations enhance ERG activity to promote tumorigenesis.
Concept:TMPRSS2–ERG may promote prostate tumorigenesis in part by outcompeting ERF at ETS consensus motifs.
Impact: Loss of ERF activity promotes ERG-dependent androgen signaling to promote prostate oncogenesis.
The TMPRSS2–ERG gene fusion drives expression of the normally silent ETS transcription factor ERG in response to androgens in approximately half of prostate cancers. Somatic point mutations in a related repressive ETS gene, ERF, have been identified in a small subset of patients with metastatic prostate cancer without the TMPRSS2–ERG fusion. Bose and colleagues investigated the role of these ERF mutations in prostate tumorigenesis. ERF mutations were found in 1%–3% of patients with prostate cancer, including loss-of-function truncating mutations and missense mutations in the ETS domain that destabilized the protein. ERF deletions were also identified, and both the mutations and deletions were largely found in TMPRSS2–ERG-negative tumors. ERG is not expressed in normal prostate epithelium, and induces prostate tumorigenesis in Pten−/− mice; conversely, ERF is expressed in normal prostate, and ERF depletion promoted prostate tumorigenesis in Pten−/− mice. Further, RNA sequencing revealed that ERF depletion induced a downregulation of basal signature genes and upregulation of genes upregulated by androgen in prostate cancer, altogether suggesting that ERF depletion phenocopies ERG overexpression. In normal human prostate and primary prostate tumors, ERF expression was inversely correlated with androgen transcriptional activity signatures, even in TMPRSS2–ERG+ cells. Mechanistically, ERF and ERG bound competitively to the ETS consensus motif; thus, when ERG expression was low, ERF chromatin binding increased and androgen-induced gene expression was suppressed. Similarly, ERG overexpression reduced ERF chromatin occupancy. Moreover, in vivo depletion of ERF accelerated the growth of ERG-dependent tumors, and induction of ERF expression prevented tumor formation and blocked androgen-dependent gene expression. Collectively, these findings support a model in which loss of ERF activity, via loss-of-function mutations or competition with TMPRSS2–ERG, promotes ERG-dependent androgen signaling and prostate cancer, and these findings may be relevant to other ETS-driven malignancies.